Observations on the Biology and Structure of three Dry Tropical Forests in South India

Tropical dry deciduous forests are found in variable climates characterized by low rainfall where woody plants possess several functional traits that permit them to endure severe water stress for several months of the year. We present an assessment of species-rich Dry Tropical Forests of the South Indian Deccan Plateau based on three large, tree-mapped field plots located in close vicinity to each other. The study includes a descriptive section (details of 130 woody plant species) and a modeling section. The modeling section presents specific species-area relations, species abundance distributions and relationships between biological attributes and individual-based structural attributes. The Monod equation is found most suitable for modeling the species-area relations confirming previous studies. The shape of the species abundance distribution follows the Weibull model which represents an alternative to the traditional lognormal model; the Weibull parameters are related linearly to species richness which is a new finding.


Introduction
Skilful and continuous observation provides essential evidence about dynamic processes in forest ecosystems (Sagarin and Pauchard, 2012). Examples of observational infrastructures are national forest inventories, field experiments and long-term observational studies. The objective of a national forest inventory is to prepare reports about the state of the forest resource at a given time and within a specific geographical context (Alekseev et al., 2019;Zeng et al., 2015). Field experiments are established to evaluate ecosystem response to specific treatments. Examples of manipulated field experiments include thinning and fertilizer treatments, short-rotation coppice trials and biodiversity-ecosystem functioning experiments.
Forest observational studies complement forest inventories, and represent an important alternative to designed experiments (Condit, 2008). The system is not manipulated, trees are mapped, and field plots are large to capture effects of scale. Once established, re-measurements continue on the same site to assess the response to environmental change and human impact. A Forest Observational Study thus represents an important research infrastructure which provides a continuous flow of information about forest ecosystem response to disturbance and changing environmental conditions. More recently, such studies, established in natural and semi-natural forests in China, India, Africa and Mexico, have emerged as an important ecological infrastructure, complementing national forest inventories and designed experiments (Gadow et al., 2016). We present three examples of such studies established in the tropical dry deciduous forests of Southern India.

Tropical Dry Deciduous Forests of India
Forest classification systems were developed by experienced ecologists who were able to see differences and commonalities without getting bogged down by the dazzling variation in small-scale site conditions, community structures and species compositions. An example is the "official" forest classification system of India (Champion and Seth, 1968 Hui and Richardson (2017) have shown how humans have been rearranging the world's biota following the era of European colonization, and especially through the post-second World War globalization. Humans have generated widespread biological invasions, leading to radical alterations to the functioning of ecosystems. Acacia species that are commercially important have been extensively planted in areas outside their natural range. Eucalypts have had modest invasive success despite their wide dissemination.
Casuarinas have only recently been widely planted and little is known of their invasion ecology (Richardson et al., 2014). The primary forests of the world have not only been modified by human-mediated introductions of organisms to areas well outside their potential range; they have also been subject to extensive use and modification by unsustainable exploitation, illegal logging (Kleinschmit et al., 2016) or sophisticated "near-natural" management (Schütz et al., 2012). Marris (2013) proclaims such new ecosystems the "new normal" and calls on scientists and the public to embrace them and to "shake off the yoke of history".
Tropical dry deciduous forests are found in severe and extremely variable climates characterized by low rainfall and nutrient-poor soils where woody plants possess several functional traits that permit them to endure severe water stress for several months of the year. Canopy trees on drying soils typically respond to an extended drought by shedding their leaves (Borchert 1994). The lack of precipitation during several months of the year produces ecosystems that have adaptated to survive a prolonged dry season.
Deciduousness is the single most important adaptation among plants to the extended droughts. According to Singh and Chaturvedi (2017), these forests are among the most vulnerable and fragmented ecosystems in the world. In India, tropical dry deciduous forests are widely distributed over a large area. The tropical dry deciduous forests of the South Indian Deccan Plateau are represented by a few protected remnants of very particular ecosystems within densely populated areas. Fig. 1 shows the distribution of forest types prepared by the Forest Survey of India, and based on the classification by Champion and Seth (1968). Improved knowledge about the structure and dynamics of these forests, based on carefully selected observational field plots with mapped trees, will contribute to improved understanding and more effective conservation of this unique natural heritage. Three such protected areas, represent the empirical basis for this study.

Study Objectives
Studies of small-scale structural patterns in forests are still relatively rare (Hui et al., 1998;Aguirre et al., 2003;Pommerening and Grabarnik, 2019). Accordingly, the objective of this study is to contribute to improved understanding of these unique ecosystems by analyzing their biology and structure in some detail. Particular objectives are: 1. to apply specific methods of structural analysis based on the attributes of neighborhoods in the vicinity of individual trees; 2. to study relationships between the relative frequency and mean size of individual trees and the structural pattern in their immediate neighborhood; 3. to assess the biological and structural differences among the three study areas by comparing the species-area relationships and species abundance distributions. We expect that our study will contribute to improved understanding of these communities, raise awareness among the scientific community, and thus contribute to more effective conservation of these unique forests in South India. The methodological approaches, described in the 'Methods' section, are applied to the observations collected. The results are discussed and compared with the results of similar studies in other regions of the world.

Descriptive Methods
Descriptive details for each woody plant species will be developed, separately for each plot, including plant taxonomies, taxonomic ratios and number of introduced or invasive species. We will also present a brief summary of the 22 species of climbers encountered in the three study areas. The climbers are classified based on the mode of climbing as twiners, lianas and stragglers.

The Species-Area Relationship
To allow comparisons of the study sites in terms of species richness, it is necessary to develop a relation between the contiguous plot area and the number of species in each plot. A number of models have been proposed to describe the species-area relation (Monod, 1950;de Caprariis et al.,1976;Gitay et al.,1991;Williams, 1995;Tjørve, 2003). Asymptotic functions are appropriate in very large plots where all species are likely to be captured by the samples. The power function is more suitable for small plot sizes where the maximum number of species is unknown. The following function, proposed by Monod (1950), represents a suitable compromise and will be used in our study to estimate the species-area relation: where a, b are empirical parameters; S is number of species; A is a contiguous forest area (m 2 ). We derive such a species-area relation by assigning sample plots of increasing size to random positions within the study area. The sampled area and associated number of species are used to derive a species-area relation (SAR) for the whole plot. Eq. 1 has the following properies: 1) when A=0, then S=0; 2) S increases with increasing A, until an asymptotic value (S max ) is reached; 3) the estimated maximum number of tree species equals a/b, which is a useful property.

The Species-Abundance Distribution
The species abundance distribution (SAD) describes the abundances of all species recorded within a forest community of interest. The SAD may explain processes of community assembly, and is believed to be one of the most ubiquitous patterns in ecology . We estimate the SAD using the Weibull distribution: where LN is the estimated logarithm of the number of individuals; k is the log (number of individuals of species 1, i.e. the species with the maximum number of individuals). SR is species rank; b and c are estimated parameters.

Forest Structure: cell-based
Information about ecosystem structure presents a useful complement to the biological analysis of species richness and abundance patterns. A first approach to characterizing structure is to subdivide a study area into smaller cells (or quadrats). The subdivision into smaller spatial units facilitates detailed analysis of small scale patterns, as well as comparison among different study areas. Fig. 3

CE: Clark & Evans index in cells
The absolute discrepancy between the distributions of these variable in two study areas was calculated using the following criterion: where p i and q i are the relative frequencies in the i'th frequency class of a particular variable in pairs of study areas that we wish to compare. The absolute discrepancy d thus represents the proportion of a particular frequency in one study area that has to be changed such that both distributions (ordered by specific frquency classes) are identical. Our analysis is limited to 20 x 20m cells.

Forest Structure: Individual-based
Forest structure may also be characterized by evaluating the immediate neighborhood of selected tree species. We will use the variables Mingling, Dominance and Size Differentiation to describe the specific neighborhood constellations of each individual species. Three measures of species-specific structural diversity are defined as follows (Gadow, 1993;c.f. Pommerening et al., 2020): Mean heterospecific fraction of trees among the k nearest neighbours of a given tree i.
Mean fraction of n nearest neighbors with a dbh <(dbh of the reference tree).
Mean of the ratio of smaller and larger tree sizes u of the k nearest neighbours subtracted from one.
The three variables represent a system for characterizing structural patterns at high resolution in a consistent set where all the variables assume values in the interval [0,1]. Mingling defines the degree of spatial segregation of the tree species (Gadow 1993;Aguirre et al., 2003;Pommerening and Grabarnik, 2019). Dominance measures the size dominance of the reference tree in relation to its immediate surrounding (Hui et al., 1998). Size Differentiation measures the variation of tree sizes between the reference tree and its nearest neighbors (Pommerening et al., 2020). Instead of selecting a particular reference species (as was done in this study), we may wish to select all trees that belong to a particular family, or all dominant trees of a given species, as reference trees with the aim to study their particular neighborhoods.
Note that reference trees located close to the plot edge may produce a biased estimate of the neighborhood constellation because some of the real nearest neighbors may be located outside, beyond the plot perimeter.
To avoid such bias, edge correction has to be employed. The simplest method involves a definition of a buffer around the plot edges. Edge correction, ensuring that the distance to the plot boundary of each reference tree must be greater than the distance to its 4th neighbor, is applied in this study to avoid biased estimates of the neighborhood parameters.

Results
We present descriptive and modeling results. The descriptive results are assumed to be of interest to biologists interested in taxonomic detail, while the modeling results allow comparisons among forest structure in a wider context.

Woody Plant Species
Appendix 1 presents a table with details for each species, separately for each plot. The information includes parameters that are assessed in routine forest inventories (mean dbh, trees per ha) as well as the means of the structural parameters Mingling (M), Dominance (D) and Size Differentiation (T). A summary of the details in Appendix 1 is presented in Tab. 2.
Almost 70 percent of all the tree and shrub species encountered in the three study sites occur on all three sites. Thirty one percent,(181-137) of 130, of species are not common to all three sites.

Neighborhood-based Results
Each tree species is characterized by an average diameter at breast height (dbh), a specific contribution to the total density of woody plants, and by a species-specific neighborhood constellation.

Cell-based Results
The absolute discrepancies among the three study areas, based on the seven cell variables, are presented in Tab. 4. Bugarikallu and Thalewoodhouse differ most in density (BA_ha), quadratic mean dbh (Dq) and diameter coefficient of variation (CVD). Bugarikallu and Doresanipalya differ in terms of species richness and diameter coefficient of variation. Doresanipalya and Thalewoodhouse differ most by density, richness, diameter coefficient of variation and cell mingling.

The Species-Area Relation
The estimated parameters a and b of the Monod model, and the graphs of the fitted functions for each field plot, are presented in Tab. 5.

The Species Abundance Distribution
Species abundance curves provide information about how communities differ in the way they are organized. The species abundance distribution generally takes a curve shape that is defined by many rare species and a few common ones (McGill et al., 2007;. Fig. 4

Discussion
A large proportion of the world's population relies directly on forests for livelihood. Sustaining these ecosystems is thus often a matter of survival. Humans have to accept the premise that the allegorical "Garden of Eden" is a dreamworld. Human impact on our planet is so overwhelming that the current period in the Earth's history has been named Anthropocene -the age of humans. Crutzen (2002) concluded that mankind has been and will remain a major environmental force in the future. In this study, we present some examples of evaluating the biology, density and structure of complex forest ecosystems. These methods are part of a scientific toolbox that enables us to make value judgments, and to choose between alternative courses of action.
Planted forests attracted much interest during the 1960's and 1970's of the past century, our study areas are self-regenerating forests. Self-regenerating forests are a vast resource. They include complex primary forests as well as exploited and degraded ecosystems, forests subject to sophisticated selection management systems, or communities dominated by invasive species. New initiatives are required to establish particular

Tree Species Richness and Diversity
In community ecology there has been more attention paid to the measurement of species diversity than to almost any other parameter. Accordingly, there is a rich literature on diversity, with many contradictory recommendations (Hubálek, 2000). Most popular are Hill's numbers as easily interpreted measures of diversity. This includes the exponential form of the Shannon 2 function (Hill's N1) and the reciprocal of Simpson's index (Hill's N2). The choice depends on whether more weight is given to the rare species (N1) or to the common species (N2).

Traditional indices of biodiversity incorporate only the numbers of species and their frequencies without
considering the biological differences among the species. Ganeshaiah et al. (1997) proposed a measure of community diversity known as the "Avalanche index". The Avalanche measure, recently "revived" by Hao et al. (2019), is defined as follows: where S is the total number of tree species, p i and p j are the relative frequencies of species i and j in the community, and d ij is a measure of the taxonomic distance between species i and j. The Avalanche diversity does not only account for the number of species and their frequencies, but also considers the taxonomic hierarchy. The Shannon entropy would be the same for two communities A and B if both have the same number of species occurring with the same frequencies. The Avalanche diversity in B would exceed that in A if the number of genera would be greater in B than in A, because the Avalanche index captures the intracommunity biological variation. The Avalanche is not only useful as an index of diversity, but also as a measure that can be used to assess the dissimilarity of two forest communities (Hao et al., 2019; see also Talents et al., 2005). Tab. 6 shows that the Hill D1 (the exponent of the Shannon index) is almost identical for Bugarikallu and Thalewoodhouse although Bugarikallu has more species than Thalewoodhouse but a lower evenness. This result is supported by the fact that the Avalanche index for Thalewoodhouse exceeds that for Bugarikallu. The number of species per family is 2.17 in Bugarikallu, and 2.46 in Thalewoodhouse (Tab. 2) which explains the slightly greater Avalanche value for Thalewoodhouse.  1962;May, 1975;.
A quantity of considerable practical relevance is the minimum contiguous area required to capture all the species within a particular region. Gadow and Hui (2007) found a relationship, based on tree-mapped field plots assessed in various regions of the world, between the maximum number of tree species within a forest region (S max , which is often known), and the minimum contiguous area required to capture all the species within that region (A min , measured in m 2 ). The minimum contiguous area was estimated in their study by the function A min = 487.8×S max 0.524 . This result implies that, for contiguous forest areas, the form of the species-area relationship is directly defined by the observed species abundance and the maximum number of species in the region. Assuming that the maximum number of species in the region around Bengaluru is 130, the estimate of the minimum contiguous area to capture all species would be 487.8*130^0.524=6250.98m 2 in each of the three plots. This area is less than that of the study areas (10000m 2 ), but inspection of the graphs of the SAR functions reveals that the estimate is quite reasonable.

Specific Relationships
The relationships between variables that are often assessed in routine forest inventories (mean dbh and number of trees), and neighborhood parameters (Mingling and Dominance) are shown in Fig. 5 for the three study areas. The relation between the number of trees per ha and the mean neighborhood mingling is estimated using a power function. The relation between the mean dbh (cm) and the mean neighborhood Dominance is estimated using the Monod function. Figure 5 presents the graphs and the equations of these relationships for each study area. The structural parameters provide additional information about the close-range neighborhood of each species. Not surprisingly, high correlation values are found between tree density and mingling for individual species. No relation was found regarding the dbh differentiation (T). Tree size variation within neighborhood groups was independent of tree size and the degree of species mingling.

Conclusions
Forests include a great variety of ecosystems where plants and animals interact with their physical environment. The challenge is to sustain their ability to function, to adapt to changing climates, and to satisfy a variety of human needs. Permanent field plots with mapped trees provide an essential "green infrastructure" for observing the dynamic evolution of such ecosystems. This study presents three abundance distribution. Non-linear relationships were established between close-range neighborhood structure, species abundance and relative dominance. A previously published general assumption for estimating the minimum contiguous area required to capture all species based on regional species richness was confirmed.

Ethics approval and consent to participate
Not applicable

Consent for publication
Not applicable

Availability of data and material
The datasets used and/or analysed during the current study are available from the corresponding author on request.

Competing interest
The authors declare that they have no competing interests.

Funding
The financial support was provided by the Department of Science & Technology (DST), Government of India under the Stragetic Programs, Large Initiatives and Coordinated Action Enabler (SPLICE) project.

Authors contributions
All authors contributed equally; KVG analyzed and interpreted the data. All authors read & approved the final manuscript.

Acknowledgements
We thank the Karnataka Forest Department (KFD) for giving us permission to establish the permanent preservation plot (PPP) in Bannerghatta National Park (BNP) and also providing us help/manpower during our studies in this elephant prone area. We are indebted to all the BNP forest watchers for accompanying us during field visits. Our heartfelt thanks are due to Ms Saswathi Misra, Ex Director, EMPRI for assistance in initiating the project works, Mr Dattaraja, IISc and Mr Kiranraddi, EMPRI for the helps in establishing the plots in forests. We also thank Mr. Suresh, , Forest Survey of India (FSI) & Mr. Abhilash, , National Institute of Hydrology for the preparation of maps of forest types and study area.
The support provided by researchers in the Centre for Climate Change, EMPRI is also gratefully acknowledged .